The Pelli-Robson Chart
Instructions
Setting Up
i. The Chart. These instructions accompany a chart
as well as scoring sheets (PDF Available). The
chart and scoring sheets are printed for all charts.
The different charts have different letter
sequences but are otherwise identical. The
letters on the chart are organized into groups of
three, i.e. “triplets”, there being two triplets per
line. Within each triplet all letters have the same
contrast. The contrast decreases from one triplet
to the next. The division into triplets is indicated
on the scoring sheet- but not on the chart itself-
by an extra-large gap. (Unlike an acuity chart, in
which the difficulty increases in the middle of
each line as well.)
ii. Mounting the chart. The chart should be hung so
that its center will be approximately at the level
of the patient’s eyes.
iii. Illuminating the chart. The chart should be
illuminated as uniformly as possible, so that the
luminance of the white areas is about 85 cd/m2.
(The acceptable range is 60 to 120 cd/m2, which
corresponds to an incident illumination of 240-
480 lumens per square meter (lux)) Avoid glare.
The patient should not see the lamps themselves
or any mirror-like reflections of the lamps on the
chart’s surface.
Contrast Sensitivity Testing
1. Testing a patient. Test patients before dilating
their pupils or applying any other drugs to their
eyes. The patient should sit or stand directly in
front of the chart so that the distance from the
eyes to the chart is about 1 meter, or 40 inches.
(The acceptable range is 34 to 46 inches.)
Patients should wear their best distance
correction and, if necessary, an additional +0.75
diopters for the 1 meter distance. (The patient’s
Copyright 2013© Denis Pelli and John Robson
Manufactured by Precision Vision, Inc. – Woodstock, IL 60098 USA 815.223.2022 – precision-vision.com
sensitivity will be unaffected by small refractive
errors because the letters are large.)
2. Recording the patient’s performance. Fill in the
patient’s name, the date and the examiner’s
name on the scoring sheet. The patient should
make a single attempt to name each letter on the
chart, starting with the dark letters in the upper
left-hand corner and reading horizontal across
the entire line. On the scoring sheet, underline or
circle each letter read correctly, strike through
any letter read incorrectly.
3. Don’t let the patient give up too soon. Patients
should be made to guess even when they believe
that the letters are invisible. You should allow
several seconds for the faintest letters to appear,
but don’t let the patient give up until he or she
has guessed incorrectly 2 of the 3 letters in a
triplet. The reliability of the results depends on
this.
4. Scoring the test. The patient’s sensitivity is
indicated by the faintest triplet for which 2 of the
3 letters are named correctly. The log contrast
sensitivity for this triplet is given by the number
on the scoring sheet nearest to the triplet. The
number may be to the right or the left of the
triplet; use the one nearest to the triplet. Enter
this number as the Log Contrast Sensitivity.
5. Testing the other eye. The patient should be
tested three times: each eye separately and both
eyes together. When testing one eye the other
eye should be covered. The three measurements
should take no more than 8 minutes in all.
Binocular log contrast sensitivity is normally 0.15
higher than monocular.
 Notes
a) The calibration. Printing processes have unavoidable
variations which make it impossible to consistently
produce contrast-sensitivity charts which meet the
standards required of a clinical instrument. Therefore,
each letter on every Pelli-Robson chart is
photometrically calibrated (by an extremely precise
reflection densitometer developed by Robson and Pelli)
to determine whether its contrast is within
specification. Charts that fail this test are discarded.
Your chart has a serial number either on the front or
back which relates to its specific calibration.
b) Life and care of the chart. The chart’s substrate and
special ink were chosen for their great stability, and
should not change the chart’s contrast significantly
within the first five years of normal use. This chart
should be replaced if it’s marred by visible marks, e.g.
fingerprints. The chart should not be cleaned or gotten
wet under any circumstances.
c) Explaining the test. This test will be unfamiliar to most
of your patients, and they may cooperate more readily
if they understand why it is being performed. Here is
one possible set of instructions: “In everyday life we
don’t just look at small black objects. Contrast
sensitivity is a more realistic assessment of how well
we see large faint objects around us. This chart is a
little different from the regular eye chart. With this
chart the letters are all uniformly large, and they face
out towards the bottom of the chart. The top line has
high-contrast letters, black on white. The letters below
them are gray and more difficult to see, very much like
looking through a fog or dirty glasses. What you must
do is read as many letters as you can. The letters at the
bottom of the chart are difficult for everyone to read,
so don’t be discouraged.” When the patient begins to
have trouble, it may be useful to provide some
strategies to help him or her make the best attempt at
seeing the letters. “Try blinking, or viewing the letter a
little eccentrically, moving your head from side to
side.” Indicate (without touching the chart) the
particular letter you want the patient to concentrate
on. “Try reading this one. Do you see something
against the white background? Is there a smudge? Is it
round or square? Does it have corners or lines you can
see? Keep trying. The whole letter may suddenly
appear to you. Go ahead and guess.”
Copyright 2013© Denis Pelli and John Robson
Manufactured by Precision Vision, Inc. – Woodstock, IL 60098 USA 815.223.2022 – precision-vision.com
d) Log contrast sensitivity. For a chart of this kind,
contrast is best defined as the difference in luminance
between the letter and background, divided by the
luminance of the background. This ratio of luminance,
known as the Weber contrast, should not be confused
with Michelson contrast, which is a different luminance
ratio commonly used for grating stimuli. The lowest
visible contrast is called the contrast threshold. It is
usual to take the reciprocal of contrast threshold to
obtain contrast sensitivity. This has the advantage that
the better a patient’s vision, the higher the sensitivity
score. Taking the base-10 logarithm of the sensitivity
has the further advantage that equal steps on this scale
correspond to equal effects. For example, looking
through a fogged window (or a cataract) which reduces
contrast by a factor of 2 would reduce the log contrast
sensitivity of any observer by 0.3, regardless of the
observer’s initial log contrast sensitivity. If it is
necessary to compute the contrast threshold c from
the log contrast sensitivity s, use the formula c = 1/10ˢ
e) Standardizing. These instructions and accompanying
chart have been designed to achieve the highest
possible comparability of results among different users.
To this end, the chart follows the luminance, font, and
letter spacing recommendations of the Committee on
Vision of the National Academy of Sciences and
National Research Council (Adv. Ophthal. 41, 103-148,
1980). Of course, this comparability will be achievable
only if the user strictly follows these instructions.
f) Accuracy. The accuracy of a single determination of a
patient’s log contrast sensitivity is determined primarily
by two factors: the accuracy of the contrast calibration
of the chart (Guaranteed to be at worst ±0.075) and
the intrinsic variability due to the probabilistic nature
of the patient’s responses, which as a result of careful
design, results in a standard deviation of only about
0.11 (see Pelli et al. 1988).
g) Further Reading. Pelli, D.G., Robson, J.G., and Wilkins,
A.J. (1988) Designing a new letter chart for measuring
contrast sensitivity. Clinical Vision Sciences2:187-199.
(Note that the article describes an early prototype
version of the chart which had a different viewing
distance.) Other articles of interest are contained in
the same issue of Clinical Vision Sciences.